
In a case that sounds more like a corporate spy thriller than a Silicon Valley personnel matter, US federal agents have arrested an OpenAI engineer and charged him with secretly installing a hidden audio recording device inside an executive conference room at the artificial intelligence company, according to Reuters.
The Justice Department said on Tuesday that Abdulrahman Alfozan, a 35-year-old Saudi national who had worked as an engineer at OpenAI since July 2025, was taken into custody over the alleged scheme. The arrest marks one of the most serious insider-security incidents ever made public at a leading AI company — and it raises fresh questions about how frontier AI labs protect their most sensitive conversations.
How the Hidden Device Was Discovered
OpenAI discovered the recording device plugged into a USB port on the back of a television in an executive conference room at one of its San Francisco buildings on September 30, 2026. Investigators believe it had been installed on or about September 28 — meaning it may have been quietly recording for up to two days before anyone noticed.
According to an FBI complaint cited by Reuters, the thumb-sized device carried a microphone and could connect to a camera. It was reportedly concealed inside toilet paper to keep it out of sight. The data stored on it was encrypted and inaccessible to investigators at the time of writing — a detail that adds a layer of mystery to what exactly was captured.
The Evidence Investigators Say They Found
Federal investigators have not yet proven their case in court, but the complaint lays out a series of allegations that investigators say point to Alfozan:
- Investigators later found a similar customized device in Alfozan's personal bag, according to the complaint.
- Surveillance footage showed him near the conference room multiple times, though he was not authorized to enter it.
- The complaint also says he used a productivity tool to ask about building addresses, meeting rooms, room capacities, calendars, and the locations of leadership and external partner meetings.
Taken together, prosecutors allege these details suggest the recording was a planned effort to capture confidential discussions, not a one-off act.

OpenAI Says Its Own AI Helped Catch Him
In a notable twist, OpenAI said in a post on X that it used its own internal AI models to help identify the person responsible and trace the device's origins. The company also said it had evidence that the employee intended to place additional recording devices in other sensitive areas of the building.
OpenAI confirmed that Alfozan was terminated and offered a measure of reassurance about what was not compromised: "At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data."
For a company whose core assets are its AI models and training data, that statement was clearly aimed at calming partners and customers even as the investigation continues.
Timeline of the Alleged Incident
- July 2025: Abdulrahman Alfozan begins working as an engineer at OpenAI.
- September 28, 2026: The hidden recording device is allegedly installed in the executive conference room.
- September 30, 2026: OpenAI discovers the device plugged into a USB port behind the room's television.
- October 2026: OpenAI refers the matter to the FBI; the investigation continues.
- October 13, 2026: The Justice Department announces the arrest and charges, per Reuters.

Why This Case Matters Beyond OpenAI
The FBI treated the referral as a matter of national security. In a statement, FBI San Francisco's special agent in charge Scott Schelble said: "Any attempt to illicitly obtain protected information from U.S. super intelligence and technology companies threatens both American innovation and the national security interests those advancements support."
The case lands at a moment when competition in artificial intelligence has become geopolitically charged. Leading AI labs now guard their research, models, and strategic discussions the way defense contractors guard weapons systems — and insider threats are among the hardest to stop.
Security experts say the episode will likely push major tech companies to tighten internal controls: more restricted access to meeting rooms, routine sweeps for unauthorized devices, and closer monitoring of who queries internal directories about leadership meetings.
What Comes Next
Alfozan has been charged, not convicted. The FBI said the investigation is continuing — which means more details, and possibly additional charges, could still emerge.
For now, the case stands as a striking reminder: in the race for AI supremacy, some of the biggest threats may come from inside the building. And even the companies building the world's most advanced AI are discovering they need it to police their own hallways.

Disclaimer: This case is ongoing and the charges are allegations; all details come from the Justice Department, FBI filings, and Reuters reporting, and may change as the investigation continues.